Estimating the value of power losses forecast in the transmission system by comparing different electricity procurement strategies

One of the main roles of the transmission system operators (TSOs) is to predict and plan the procurement of the active power losses in transmission systems. TSO can reduce operating cost using power losses forecasts and, in principle, the higher the power losses forecast accuracy is the lower are the TSO's operating cost. The paper continuous on our previous work where we describe the newly designed machine learning based tool forecasting power losses that has been implemented at the Croatian TSO. After more than a year of continuous everyday operation of the tool we now evaluate its real world performance and use its results in further analysis. To reap the benefits of accurate power losses forecast, the TSO has to choose a good strategy for procurement of those losses on the electricity markets. In this work we present an analysis of hypothetical procurement strategies. Electricity markets that consider in the analysis are Croatian day-ahead, intraday and balancing markets. The baseline procurement strategies are procurement on single markets, while advanced strategies combine all three markets for lowest cost. Special attention is given to the imbalance market, as the differences between maximum and minimum price are very large. The use case for the presented analysis is Croatian TSO, but most of the analysis should generalise well to other TSOs. Our analysis shows that the value of forecast is greatly influenced by the market rules. The value of a forecast is much larger in two-price imbalance systems as compared to one price imbalance systems.